There is a growing consensus among acade-mics, policy makers, and even politicians thatpoverty and inequality should no longer betreated as soft “social issues” that can safely besubordinated to more fundamental interestsin maximizing total economic output. Themost important sources of this newfoundconcern with poverty and inequality are (1)the spectacular increase in economic inequal-ity and other forms of disadvantage in manylate-industrial countries (the takeoff account);(2) the striking persistence of other forms ofinequality (e.g., racially segregated neighbor-hoods, occupational sex segregation) despitedecades of quite aggressive egalitarian reform(the persistence account); (3) an emerging con-cern that poverty and inequality may havenegative macro-level effects on terrorism,total economic production, and ethnic unrest(the macro-level externalities account); (4) agrowing awareness of the negative individual-level effects of poverty on health, politicalparticipation, and a host of other life condi-tions (the micro-level externalities account); (5)the rise of a “global village” in which spatialdisparities in the standard of living have be-come more widely visible and hence increas-ingly difficult to ignore (the visibility ac-count); (6) the ongoing tendency to exposeand delegitimate new types of inequalities(based on sexual orientation, disability, or cit-

izenship) that, not so long ago, were taken forgranted, rarely discussed, and barely seen (thenew inequalities account); and (7) a growingcommitment to a broader conception ofhuman entitlements that encompasses rightsto basic social amenities (e.g., housing) aswell as rights to basic forms of social partici-pation, such as employment (the social inclu-sion account).

This list is remarkable in two ways. First,only two of the seven reasons for our new-found interest in poverty and inequality areabout brute empirics (i.e., the growth or in-transigence of inequality), while all others areabout changes in how we have come to view,study, and evaluate those empirics. Whenscholars now argue, for example, that in-equality has multifarious unintended effects(i.e., an externalities account), they presum-ably don’t mean to suggest that such effectssuddenly multiplied in the contemporary pe-riod (although perhaps they have). Rather,we are to understand that inequality was al-ways rife with externalities, however inade-quately we may have appreciated them in thepast. While changes in empirics hardly ex-haust, then, the sources of our growing con-cern with inequality, this is not to gainsay theequally important point that such changes,especially the recent takeoff in income in-equality, are likely a core reason why inequal-

2 – Melvin M. Tumin

❚ David B. Grusky and Manwai C. Ku

Gloom, Doom, and Inequality

This is a commissioned chapter that draws heavily on material in previous publications (David B. Grusky andSzonja Szelényi, “The Rise and Fall of Benign Narratives About Inequality,” David B. Grusky and SzonjaSzelényi, eds., The Inequality Reader: Contemporary and Foundational Readings in Race, Class, and Gender, 2007 byWestview Press; David B. Grusky, “The Past, Present, and Future of Social Inequality,” David B. Grusky, ed., So-cial Stratification: Class, Race, and Gender in Sociological Perspective, 2nd edition by Westview Press).

0813343739-01.qxd 10/1/07 11:32 AM Page 2

ity has come to be understood as a funda-mental social problem of our time.

The above list is no less remarkable for therelatively minor role that normative concernsplay. To be sure, there appears to be a grow-ing sentiment that, at minimum, contempo-rary social systems should guarantee an op-portunity to all citizens to participate ineconomic life and hence avoid the most ex-treme forms of social and economic exclu-sion (i.e., the social inclusion account). Itwould nonetheless be a mistake to under-stand the rising interest in poverty and in-equality as principally fueled by some suddenrealization that social inclusion is a funda-mental social good. Indeed, far from treatinginequality as exclusively a moral problem,there is a growing tendency to emphasize itsprofound consequences and threats for theworld community as a whole (i.e., the macro-level externalities account). The rhetoric of“sustainability,” although more frequentlyfeatured in discussions of environmentalproblems, is increasingly taken as relevant todiscussions of inequality as well (see Mullahy,Robert, and Wolfe, Ch. 95; Stiglitz, Ch. 109;Krueger and MaleOková, Ch. 111). Inadopting this rhetoric, the claim is that ex-treme inequality is counterproductive notjust because it reduces total economic out-put, but also because other very legitimateobjectives, such as reducing mortality ratesor the threat of terrorism, might be compro-mised if inequality remains so extreme. Bythis logic, social policy must simultaneouslybe oriented to increasing economic outputand restraining the rise of debilitating andcounterproductive forms of inequality, arather more complicated maximization prob-lem than that conventionally taken on (Fis-cher et al., Ch. 5; Krueger, Ch. 6).

Core Fields of Inquiry

The foregoing is to suggest that the maintask of stratification scholarship, that of de-scribing the contours of inequality and ex-

plaining its causes, has come to be viewed asan increasingly important and central en-deavor, and not just within the narrow worldof academia. The purpose of this introduc-tory essay is to lay out how the field of strat-ification has developed, where it currentlystands, and how this volume represents andconveys that field.

It may be useful to begin such a task byspecifying the main questions around whichthe field has to this point developed. For eachof the eight subfields listed below, we providesome examples of the types of questions thathave been featured, our intent being to con-vey the prevailing style of inquiry rather thansome more exhaustive listing.

Forms and Sources of Stratification What have been the major forms of inequal-ity in human history? Can the ubiquity of in-equality be attributed to individual differ-ences in talent or ability (e.g., Fischer et al.,Ch. 5)? Is some amount of inequality an in-evitable feature of human life (e.g., Davisand Moore, Ch. 2; Tumin, Ch. 4)? Why isincome inequality increasing in so many late-industrial countries (e.g., Morris and West-ern, Ch. 7; Piketty and Saez, Ch. 8)?

Structure of Contemporary Inequality Is inequality organized into a small number ofsocial classes or occupational groups (e.g.,Grusky and Sørensen, Ch. 17)? Or does ittake on a largely gradational form featuringwholly incremental differences of income orstatus (e.g., Blau and Duncan, Ch. 19; Hauserand Warren, Ch. 24; Sen, Ch. 26)? Which ofthese two forms became more prominent withthe transition into modernity or late moder-nity (e.g., Weeden et al., Ch. 27)?

Inequality at the Extremes Do political, economic, and cultural elitescome together to form a single “upper class”(e.g., Mosca, Ch. 28; Mills, Ch. 29; Giddens,Ch. 30)? Is there likewise a well-formed “un-derclass” of the poor and dispossessed (e.g.,

Gloom, Doom, and Inequality – 3

0813343739-01.qxd 10/1/07 11:32 AM Page 3

Wilson, Ch. 37)? Are elites principally drawnfrom long-standing elite families and thepoor from long-standing poverty-strickenfamilies (e.g., Domhoff, Ch. 31)?

Generating Inequality How frequently do individuals move intonew classes, occupations, or income groups(e.g., Breen, Ch. 48)? Is the United States anespecially mobile society (e.g., Erikson andGoldthorpe, Ch. 47)? To what extent are oc-cupational outcomes determined by suchforces as intelligence, effort, schooling, aspi-rations, social contacts, and individual luck(e.g., Jencks et al., Ch. 51; Harding et al.,Ch. 52; Granovetter, Ch. 59; Fernandez andFernandez-Mateo, Ch. 62)?

Race and EthnicityWhat accounts for racial and ethnic differ-ences in grades, test scores, high school grad-uation, and college attendance (e.g., Steele,Ch. 71)? What accounts for racial and ethnicdifferences in hiring, promotion, and pay(e.g., Bertrand and Mullainathan, Ch. 70)?Why are neighborhoods so deeply segregatedby race and ethnicity (e.g., Massey and Den-ton, Ch. 38; Wilson, Ch. 76)? Why are in-terracial marriages becoming more common(e.g., Farley, Ch. 65)?

Gender InequalityWhy has the gender gap in educational at-tainment disappeared altogether in somecountries? Why has occupational sex segrega-tion proven, by contrast, to be so durable(e.g., Reskin, Ch. 86; Grusky and Levanon,Ch. 87)? What accounts for the gender paygap (e.g., Petersen and Morgan, Ch. 88;England, Ch. 89; Tam, Ch. 90)? Why hasthe historic decline in the pay gap stalled oflate (e.g., Blau and Kahn, Ch. 91)?

Consequences of InequalityDoes our class position determine our poli-tics, attitudes, and behaviors (Bourdieu, Ch.93; Chan and Goldthorpe, Ch. 96; Lareau,

Ch. 97; Hout and Moodie, Ch. 99)? Are sucheffects of class weakening (e.g., Hout andMoodie, Ch. 99)?

Future of InequalityIs a new “high-modern” form of stratificationemerging (e.g., Giddens, Ch. 106; Beck andLau, Ch. 108)? Will inequality continue togrow in many late-industrial countries (e.g.,Firebaugh, Ch. 110)? Is the concept of socialclass still useful in describing contemporaryforms of inequality (e.g., Pakulski and Wa-ters, Ch. 107)? Are inequality regimes gradu-ally shedding their distinctive features andconverging towards some common regime(e.g., Esping-Andersen, Ch. 102)?

The foregoing questions all adopt a criticalorientation to human stratification systemsthat is distinctively modern in its underpin-nings. For the greater part of human history,the existing stratification order was regarded asan immutable feature of society, and the im-plicit objective of commentators was to ex-plain or justify this order in terms of religiousor quasi-religious doctrines (e.g., Bottomore1965; Tawney 1931). It was only with the En-lightenment that a critical “rhetoric of equal-ity” emerged in opposition to the civil andlegal advantages of the aristocracy and otherprivileged status groupings. In its most radicalform, this egalitarianism led to Marxist inter-pretations of human history, and it ultimatelyprovided the intellectual underpinnings forsocialist stratification systems. While much ofstratification theory has been formulated inreaction and opposition to these early forms ofMarxist scholarship,1 the field nonethelessshares with Marxism the distinctively Enlight-enment premise that individuals are to beequally valued and respected regardless of dif-ferences in their economic or social standing(see Meyer, Ch. 105; also Tawney 1931).

Basic Concepts

The eight lines of questioning describedabove are elaborated in the remainder of this

4 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 4

essay. Before turning to such elaboration, itwill be helpful to introduce some of the basicconcepts used to describe the amount andstructure of inequality, the genesis and persis-tence of inequality, and the effects of in-equality on social behavior of all kinds.

The key components of stratification sys-tems are (1) the institutional processes thatdefine certain types of goods as valuable anddesirable, (2) the rules of allocation that dis-tribute these goods across various jobs or oc-cupations in the division of labor (e.g., doc-tor, farmer, “housewife”), and (3) themobility mechanisms that link individuals tojobs and thereby generate unequal controlover valued resources. It follows that inequal-ity is produced by two types of matchingprocesses: The social roles in society are firstmatched to “reward packages” of unequalvalue, and individual members of society arethen allocated to the positions so defined andrewarded.2 In all societies, there is a constantflux of occupational incumbents as newcom-ers enter the labor force and replace dying,retiring, or out-migrating workers, yet thepositions themselves and the reward pack-ages attached to them typically change onlygradually. As Schumpeter (1953) puts it, theoccupational structure can be seen as “a hotel. . . which is always occupied, but always bydifferent persons” (p. 171).

What types of rewards are distributed viathese two matching processes? It is increas-ingly fashionable to recognize that inequalityis “multidimensional,” that income is accord-ingly only one of many important resources,and that income redistribution in and of it-self would not eliminate inequality (e.g., Sen,Ch. 26). When a multidimensionalist ap-proach is accordingly taken, one might use-fully distinguish among the eight types of as-sets listed in Table 1, each understood asvaluable in its own right rather than a mereinvestment item.3 The assets of Table 1 mayof course serve investment as well as con-sumption functions. For example, mosteconomists regard schooling as an invest-

ment that generates future streams of in-come, while some sociologists likewise regardsocial networks as forms of capital that canbe parlayed into educational credentials, in-come, and other valued goods.4 While mostof the assets listed in Table 1 are clearly con-vertible in this fashion, they are not necessar-ily regarded as investments by the individualsinvolved. In fact, many valuable assets can besecured at birth or through childhood social-ization (e.g., the “good manners” of the aris-tocracy), and they are therefore acquiredwithout the beneficiaries explicitly weighingthe costs of acquisition against the benefits offuture returns.5

The implicit claim underlying Table 1 isthat the listed assets exhaust all possible con-sumption goods and, as such, constitute theraw materials of stratification systems. Thestratification field has developed a vocabularythat describes how these raw materials aredistributed across members of the popula-tion. The main parameters of interest pertainto (1) the overall amount of inequality, (2)the extent to which individuals are lockedpermanently into certain positions (i.e.,rigidity), (3) the extent to which ascription(e.g., racial discrimination) is used for pur-poses of allocation, and (4) the degree towhich the various dimensions of inequalitycohere (or are “crystallized”). These fourclasses of parameters are reviewed below.

The overall amount of inequality in anygiven resource (e.g., income) may be definedas its dispersion or concentration among theindividuals in the population. Althoughmany scholars seek to characterize the overalllevel of societal inequality with a single para-meter, such attempts will obviously be com-promised insofar as some types of rewards aredistributed more equally than others. Thiscomplexity clearly arises in the case of con-temporary inequality regimes. The recentdiffusion of “citizenship rights,” for example,implies that civil goods have come to bewidely dispersed across all citizens, whereaseconomic and political goods continue to be

Gloom, Doom, and Inequality – 5

0813343739-01.qxd 10/1/07 11:32 AM Page 5

disproportionately controlled by a relativelysmall elite.

The rigidity of a stratification system is in-dexed by the continuity (over time) in the so-cial standing of its members. The stratifica-tion system is highly rigid, for example, if thecurrent wealth, power, or prestige of individ-uals can be accurately predicted on the basisof their prior statuses or those of their par-ents. It should again be emphasized that suchrigidity (or “social closure”) will typicallyvary in degree across the different types of as-sets listed in Table 1.

The stratification system rests on ascriptiveprocesses to the extent that conditions presentat birth (e.g., parental wealth, sex, race) in-fluence the subsequent social standing of in-dividuals. If ascriptive processes of this sortare in operation, it is possible (but by nomeans guaranteed) that the underlying traits

themselves will become bases for group for-mation and collective action (e.g., race riots,feminist movements). In contemporary soci-eties, ascription of all kinds is usually seen asundesirable or discriminatory, and muchgovernmental policy (e.g., antidiscriminationlegislation) is therefore directed toward fash-ioning a stratification system in which indi-viduals acquire resources solely by virtue oftheir achievements.6

The degree of crystallization is indexed bythe correlations among the various resources(e.g., income, wealth, education) that are so-cially valued. If these correlations are strong,then the same individuals (i.e., the “upperclass”) will consistently appear at the top ofall hierarchies, while other individuals (i.e.,the “lower class”) will consistently appear atthe bottom of the stratification system. Bycontrast, various types of status inconsisten-

6 – Melvin M. Tumin

Table 1. Types of Assets and Examples of Advantaged and Disadvantaged Groups

Assets Examples

Asset group Examples of types Advantaged Disadvantaged

1. Economic Wealth Billionaire Bankrupt workerIncome Professional Laborer Ownership Capitalist Worker (i.e., employed)

2. Power Political power Prime minister Disenfranchised personWorkplace authority Manager Subordinate workerHousehold authority Head of household Child

3. Cultural Knowledge Intelligentsia UneducatedDigital culture Silicon Valley resident Residents of other places“Good” manners Aristocracy Commoner

4. Social Social clubs Country club member NonmemberWorkplace associations Union member NonmemberInformal networks Washington A-list Social unknown

5. Honorific Occupational Judge Garbage collectorReligious Saint ExcommunicateMerit-based Nobel Prize winner Nonwinner

6. Civil Right to work Citizen Illegal immigrantDue process Citizen Suspected terrorist Franchise Citizen Felon

7. Human On-the-job Experienced worker Inexperienced workerGeneral schooling College graduate High school dropoutVocational training Law school graduate Unskilled worker

8. Physical Mortality Person with long life A “premature” death(i.e., health) Physical disease Healthy person Person with AIDS, asthma

Mental health Healthy person Depressed, alienated

0813343739-01.qxd 10/1/07 11:32 AM Page 6

cies (e.g., a poorly educated millionaire) willemerge in stratification systems with weaklycorrelated hierarchies, and it is correspond-ingly difficult in such systems to define aunitary set of classes that have predictivepower with respect to all resources.

Measuring Inequality

Given the complexity of contemporary in-equality, one might expect most scholars toadopt a multidimensional approach to char-acterizing it, an approach that might beginby specifying the multivariate “inequalityspace” of valued goods listed in Table 1. Itnonetheless remains conventional amongsociologists to describe stratification sys-tems in terms of discrete classes or stratawhose members are presumed to have simi-lar levels or types of assets. In the most ex-treme versions of this approach, the result-ing classes are assumed to be real entitiesthat preexist the distribution of assets, andmany scholars therefore refer to the “effects”of class location on the assets that their in-cumbents control. We present here a styl-ized history of such class models as well ascompeting approaches to characterizing thestructure of inequality.

The claim that inequality takes on a “classform” is one of the few distinctively sociolog-ical contributions to inequality measurementand stands as the main alternative to ap-proaches that either focus exclusively on in-come inequality or analyze the many dimen-sions of inequality independently andseparately. The main advantage of class-basedmeasurement, as argued by sociologists, isthat conventional class categories (e.g., pro-fessional, manager, clerk, craft worker, la-borer, farmer) are institutionalized within thelabor market and are accordingly more thanpurely nominal or statistical constructions.The labor market, far from being a seamlessand continuous distribution of incomes, isinstead understood as a deeply lumpy entity,with such lumpiness mainly taking the form

of institutionalized groups (i.e., “classes”)that constitute prepackaged combinations ofthe valued goods listed in Table 1.

Within sociology, the implicit critique,then, of income-based approaches rests notso much on the argument that the incomedistribution is just one of many distributionsof interest (i.e., multidimensionalism), butrather on the argument that measurementstrategies based on the income distributionalone impose an excessively abstract, ana-lytic, and statistical lens on a social worldthat has much institutionalized structure toit. This structure takes the tripartite form ofa set of social classes that are privileged undercapitalist labor markets (e.g., capitalists, pro-fessionals, managers), a set of social classesthat are less privileged under advanced capi-talism (e.g., routine nonmanuals, craft work-ers, operatives), and an “underclass” thatstands largely outside the labor market and isaccordingly deeply disadvantaged in marketsystems. The rise of class models shouldtherefore be understood as a distinctively so-ciological reaction to the individualism ofthe income paradigm and other unidimen-sional approaches to measuring inequality.

The foregoing account, which is a largelyconsensual rendition of the rationale for so-cial class measurement, nonetheless concealsmuch internal debate within the field on howbest to identify and characterize the bound-aries dividing the population into classes. Wereview below the three main phases in whichthese debates have played out.

Structuralist Phase (ca. 1945–1985)The class models of the postwar period restedimplicitly or explicitly on the assumption thatclasses are bundles of endowments (e.g., edu-cation levels), working conditions (e.g.,amount of autonomy), and reward packages(e.g., income) that tend to cohere together(esp. Giddens, Ch. 14; Parkin, Ch. 15). Themiddle class of “craft workers,” for example,comprises individuals with moderate educa-tional investments (i.e., secondary school cre-

Gloom, Doom, and Inequality – 7

0813343739-01.qxd 10/1/07 11:32 AM Page 7

dentials), considerable occupation-specific in-vestments in human capital (i.e., on-the-jobtraining), average income coupled with sub-stantial job security (at least until deindustri-alization), middling social honor and pres-tige, quite limited authority and autonomyon the job, and comparatively good healthoutcomes (by virtue of union-sponsoredhealth benefits and regulation of workingconditions). By contrast, the underclass maybe understood as comprising a rather differ-ent package of conditions, a package thatcombines minimal educational investments(i.e., secondary school dropouts), limited op-portunities for on-the-job training, result-ingly intermittent labor force participationand low income, virtually no opportunitiesfor authority or autonomy on the job (duringthose brief bouts of employment), relativelypoor health (by virtue of lifestyle choices andinadequate health care), and much social den-igration and exclusion. The other classes ap-pearing in conventional class schemes (e.g.,professional, managerial, routine nonmanual)may likewise be understood as particularcombinations of scores on the dimensions ofTable 1.

For the purposes of illustration, consider asimplified case in which the multidimen-sional “inequality space” comprises onlythree individual-level variables (e.g., educa-tion, autonomy, income), thus allowing theclass hypothesis to be readily graphed. Addi-tionally, assume that the class structure canbe represented by six classes (e.g., profes-sional, managerial, sales and clerical, craft, la-borer, farm), signified in Figure 1 by six dif-ferent symbols (dark squares, light squares,dark circles, etc.). As shown in this figure, thetwo main claims underlying the class hy-pothesis are that (1) the structural conditionsof interest tend to cluster together into char-acteristic packages, and (2) these packages ofconditions correspond to occupational oremployment groupings. For structuralists,the inequality space is presumed to have arelatively low dimensionality, indeed a di-

mensionality neither more nor less than thenumber of postulated classes. The individu-als falling within the classes constituting thisscheme will accordingly have endowments,working conditions, and reward packagesthat are close to the averages prevailing fortheir classes. Moreover, even when individualscores deviate from class averages, the con-ventional class-analytic assumption (albeitwholly untested) is that the contextual effectof the class is dominant and overcomes anysuch individual-level deviations. For exam-ple, a full professor who lacks a Ph.D. is pre-sumably just as marketable as a fully creden-tialed (but otherwise comparable) fullprofessor, precisely because membership inthe professorial class is a master status thattends to dominate all others.

The postwar period also was notable for aflourishing of gradational measurement ap-proaches that again treated occupations as thefundamental units of analysis but assumedthat such occupations may be ordered into aunidimensional socioeconomic or prestigescale (e.g., Treiman, Ch. 20; Goldthorpe andHope, Ch. 21). In Figure 1, we assumed thatthe class structure cannot be understood insimple gradational terms, meaning that atleast some classes were formed by combininghigh values on one dimension with low valueson another. It is possible, however, that thestructural conditions of interest tend to co-

8 – Melvin M. Tumin

Figure 1. Class Regime

0813343739-01.qxd 10/1/07 11:32 AM Page 8

vary linearly, thus generating a class structureof the very simple type represented in Figure2. In a regime of this sort, inequality becomesrather stark (i.e., crystallized), as privilege onone dimension implies very reliably privilegeon another. There should accordingly bemuch interest in determining whether in-equality indeed takes this form. Unfortu-nately, inequality scholars of the postwar pe-riod didn’t typically test the crystallizationassumption, but rather simply assumed that itheld and proceeded to develop socioeco-nomic scales that treated education and in-come as the main dimensions of interest (andranked occupations by averaging scores onthese two dimensions).

It may be noted that many neo-Marxianscholars during this period also deviatedfrom a strict multidimensional stance bynominating particular dimensions withinTable 1 as being theoretically crucial andhence the appropriate basis upon which so-cial classes might be defined. There are nearlyas many claims of this sort as there are di-mensions in Table 1. To be sure, Marx ismost commonly criticized for placing “al-most exclusive emphasis on economic factorsas determinants of social class” (Lipset 1968,p. 300), but in fact much of what passed forstratification theorizing during this periodamounted to reductionism of some kind, al-beit often an expanded version of reduction-

ism in which two or three dimensions werenominated as especially crucial. When a re-ductionist position is adopted, the rationalefor a class model is not typically that classesare coherent packages of conditions (as rep-resented in Figure 1), but rather that thenominated dimension or dimensions are cru-cial in defining interests and will accordinglycome to be the main sources of social action.The classic Marxian model, for example, hasworkers ultimately appreciating that theirstatus as workers (i.e., nonowners) definestheir interests.

Culturalist Phase (ca. 1985–1995)In the mid-1980s, Bourdieu (Ch. 93) andother sociologists (esp. Wilson, Ch. 36)sought to develop a culturalist rationale forclass models, a rationale that rested on theclaim that classes are not merely constella-tions of structural conditions (e.g., workingconditions, rewards) but are also sociallyclosed groupings in which distinctive cul-tures emerge and come to influence atti-tudes, behaviors, or even preferences of classmembers. Throughout this period, many so-ciologists continued to work with more nar-rowly structuralist definitions of class (e.g.,Wright, Ch. 11; Erikson and Goldthorpe,Ch. 47), but Bourdieu (Ch. 93) and Wilson(Ch. 37) were instrumental in pressing theclaim that class-specific cultures are a defin-ing feature of inequality systems (also,Lareau, Ch. 97).

The two main forms of closure that serveto generate class-specific cultures are residen-tial segregation (e.g., urban ghettos) andworkplace segregation (e.g., occupational as-sociations). As Wilson notes, members of theunderclass live in urban ghettos that are spa-tially isolated from mainstream culture, thusallowing a distinctively oppositional cultureto emerge and reproduce itself. The effects ofresidential segregation operate, by contrast,in more attenuated form for other socialclasses; after all, residential communitiesmap only imperfectly onto class categories

Gloom, Doom, and Inequality – 9

Figure 2. Gradational Regime

0813343739-01.qxd 10/1/07 11:32 AM Page 9

(i.e., the demise of the “company town”),and social interaction within contemporaryresidential communities is in any event quitesuperficial and cannot be counted upon togenerate much in the way of meaningful cul-ture. If distinctive cultures emerge outsidethe underclass, they do so principallythrough the tendency for members of thesame occupation to interact disproportion-ately with one another in the workplace andin leisure activities. In accounting, for exam-ple, for the humanist, antimaterialist, andotherwise left-leaning culture and lifestyle ofsociologists, class analysts would stress theforces of social closure within the workplace,especially the liberalizing effects of (1)lengthy professional training and socializa-tion into the “sociological worldview,” and(2) subsequent interaction in the workplacewith predominantly liberal colleagues.

When occupations or classes are allowed tohave cultures in this fashion, one naturallywishes to better understand the content ofthose cultures and, in particular, the relation-ship between such content and the structuralconditions (i.e., endowments, outcomes, in-stitutional setting) that a class or occupa-tional situation implies. At one extreme, classcultures may be understood as nothing morethan “rules of thumb” that encode optimiz-ing behavioral responses to prevailing institu-tional conditions, rules that allow class mem-bers to forego optimizing calculationsthemselves and rely instead on cultural pre-scriptions that provide reliable and economi-cal shortcuts to the right decision. For exam-ple, Breen and Goldthorpe (Ch. 64) arguethat working-class culture is disparaging ofeducational investments not because of somemaladaptive oppositional culture, but be-cause such investments expose the workingclass (more so than other classes) to a real riskof downward mobility. In most cases, work-ing-class children lack insurance in the formof substantial family income or wealth,meaning that they cannot easily recover froman educational investment gone awry (i.e.,

dropping out); and those who nonethelessundertake such an investment therefore facethe real possibility of substantial downwardmobility. The emergence, then, of a working-class culture that regards educational invest-ments as frivolous may be understood as en-coding that conclusion and thus allowingworking-class children to undertake optimiz-ing behaviors without explicitly engaging indecision tree calculations. The behaviors thata “rule of thumb” culture encourages are,then, deeply adaptive because they take intoaccount the endowments and institutionalrealities that class situations encompass (alsosee Morgan, Ch. 55).

The foregoing example may be understoodas one in which a class-specific culture in-structs recipients about appropriate (i.e., op-timizing) means for achieving ends that arewidely pursued by all classes. Indeed, theprior “rule-of-thumb” account assumes thatmembers of the working class share the con-ventional interest in maximizing labor mar-ket outcomes, with their class-specific culturemerely instructing them about the approachthat is best pursued in achieving that con-ventional objective. At the other extreme,one finds class-analytic formulations thatrepresent class cultures as more overarchingworldviews, ones that instruct not merelyabout the proper means to achieve ends butadditionally about the proper valuation ofthe ends themselves. For example, some classcultures (e.g., aristocratic ones) place an es-pecially high valuation on leisure, with mar-ket work disparaged as “common” or “pollut-ing” (Veblen, Ch. 92). This orientationpresumably translates into a high reservationwage within the aristocratic class. Similarly,“oppositional cultures” within the underclassmay be understood as worldviews that placean especially high valuation on preserving re-spect and dignity for class members, with ofcourse the further prescription that theseends are best achieved by (1) withdrawingfrom and opposing conventional mainstreampursuits, (2) representing conventional mo-

10 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 10

bility mechanisms (e.g., higher education) astailor-made for the middle class and, by con-trast, unworkable for the underclass, and (3)pursuing dignity and respect through othermeans, most notably total withdrawal fromand disparagement of mainstream pursuits.This is a culture, then, that advocates that re-spect and dignity deserve an especiallyprominent place in the utility function andthat further specifies how those ends mightbe achieved.

It should by now be clear that sociologistsoperating within the class-analytic traditionhave adopted very strong assumptions abouthow inequality and poverty are structured.As noted, intrinsic to the class concept aresuch claims as (1) the space of outcomes andcapabilities has a (low) dimensionality equal-ing the number of social classes, (2) the classlocations of individuals become master sta-tuses that dominate (or at least supplement)the effects of individual-level endowments,and (3) such class locations are socially closedand come to be associated with adaptive ormaladaptive cultures. The prior claims havebeen unstated articles of faith among classanalysts in particular and sociologists moregenerally. In this sense, class analysts have be-haved rather like stereotypical economists,the latter frequently being parodied for theirwillingness to assume most anything pro-vided that it leads to an elegant model.

Postmodernist Phase (ca. 1995–Present Day)The third phase of conceptual work withinsociology has been marked, however, by anincreased willingness to challenge the as-sumptions underlying the class-analytic sta-tus quo. In recent years, such criticisms ofthe class-analytic enterprise have escalated,with many postmodernist scholars now feel-ing sufficiently emboldened to argue that theconcept of class should be abandoned alto-gether. Although the postmodern literature isnotoriously fragmented, the variant of post-modernism that is most relevant here pro-

ceeds from the assumption that the labormovement is rooted in the old and increas-ingly irrelevant conflicts of industrial capital-ism, that political parties have abandonedclass-based platforms in favor of those ori-ented toward values and lifestyles, and thatclass-based identities accordingly becomeever weaker and more attenuated (see Houtand Moodie, Ch. 99, for a relevant discus-sion). The resulting “individualization of in-equality” (e.g., Pakulski and Waters, Ch.107; Beck and Lau, Ch. 108) implies thatlifestyles and consumption practices are be-coming decoupled from work identities aswell as other status group memberships. Thestratification system may be regarded, then,as a “status bizarre” (Pakulski and Waters,Ch. 107) in which identities are actively con-structed as individuals select and are shapedby their multiple statuses.7

This hypothesis, which is represented inextreme form by Figure 3, has not yet beensubjected to convincing empirical test andmay well prove to be premature. Moreover,even if lifestyles and life chances are truly de-coupling from economic class, this ought notbe misunderstood as a more general declinein inequality per se. The brute facts of in-equality will still be with us even if socialclasses of the conventional form are weaken-ing. Most obviously, income inequality isclearly on the rise (e.g., Piketty and Saez, Ch.

Gloom, Doom, and Inequality – 11

Figure 3. Disorganized Inequality

0813343739-01.qxd 10/1/07 11:32 AM Page 11

8-; Morris and Western, Ch. 7), while otherforms of inequality show no signs of wither-ing away. The postmodernist hypothesisspeaks, then, to the way in which inequalityis organized, not to the overall amount ofsuch inequality.

Analyzing Allocation

Although inequality scholars have longsought to understand how different “rewardpackages” are attached to different social po-sitions, an equally important task within thefield is that of understanding the rules bywhich individuals are allocated to the socialpositions so defined and rewarded. The lan-guage of stratification theory makes a sharpdistinction between the distribution of socialrewards (e.g., the income distribution) andthe distribution of opportunities for securingthese rewards. As sociologists have frequentlynoted, it is the latter distribution that gov-erns popular judgments about the legitimacyof stratification: The typical American, forexample, is quite willing to tolerate substan-tial inequalities in power, wealth, or prestigeprovided that the opportunities for securingthese social goods are distributed equally. Ifthe competition has been fairly run, we arequite willing to reward the winners and pun-ish the losers.

The study of opportunities (or “capabili-ties”) is no less fashionable among econo-mists. The main motivation, however,among economists for studying opportuni-ties is not some intrinsic interest in mobilityprocesses themselves, but rather a concernthat standard outcome-based measures of in-equality are tainted by the confounding ef-fect of differential tastes. For example, anemployee with a well-developed taste forleisure will presumably opt to work for rela-tively few hours, leading to low earnings butnonetheless optimal utility (by virtue of thehigh valuation placed on leisure). Wheneverincome inequality is generated through theoperation of differential tastes, most econo-

mists would argue that it should be regardedas quite unproblematic, given that low-in-come workers are simply choosing, by virtueof their particular tastes, to trade off incomefor some other valuable good (e.g., leisure).This line of reasoning implies that inequalityscholars should measure the distribution ofopportunities that prevails before differentialtastes can express themselves. The main taskof an inequality scholar under this formula-tion is to determine whether “capabilities”(i.e., opportunities to secure rewards) areequally distributed, not whether rewardsthemselves, which reflect the operation oftastes, are equally distributed.

It follows that sociologists and economistshave become quite interested, albeit for dif-ferent reasons, in the study of opportunityand how it is unequally distributed. In mostof the resulting research, the liberal ideal ofan open and discrimination-free system istreated as an explicit benchmark, and theusual objective is to expose any inconsisten-cies between this ideal and the empirical dis-tribution of life chances. This objective un-derlies analyses of the gross effects of classorigins on class destinations (e.g., Feather-man and Hauser, Ch. 46; Solon, Ch. 49;Conley, Ch. 63) as well as analyses of the neteffects of gender, race, and class backgroundafter controlling education and related mea-sures of achievement or merit (e.g., Blau andDuncan, Ch. 50; Sewell, Haller, and Portes,Ch. 53). We have presented a wide range ofsuch analyses in this volume. Additionally,experimental approaches to measuring dis-crimination have recently become popular,most notably “audit studies” that proceed by(1) sending employers resumes that are iden-tical save for the applicant’s gender, race, orclass, and (2) then examining whether callback rates (for interviews) are nonethelessdifferent across such groups. Although theavailable statistical and experimental studiesall indicate that opportunities are far fromequal (e.g., Bertrand and Mullainathan, Ch.70; Pager, Ch. 72; also see Goldin and

12 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 12

Rouse, Ch. 80), there remains some debateabout whether or to what extent such in-equalities are declining or will continue todecline. We turn to such debates next.

A Compressed History of Inequality

The task of identifying the essential dynam-ics underlying social change has long beenfundamental to sociology, but nowhere isthis interest better developed or more funda-mental than within the field of inequalityanalysis. The remainder of our essay is de-voted, then, to laying out some of the mostcompelling narratives about how inequalityhas evolved. We begin this task by develop-ing a quite conventional typology of inequal-ity forms over the long span of human his-tory and then turn to the question of howthe narrative-building enterprise has changedin recent times (i.e., the last half-century).

Although stratification systems can ofcourse be defined by a great many parame-

ters, we proceed here by building a typologyof inequality regimes in terms of the parame-ters of inequality that were introduced at thebeginning of this essay (i.e., amount of in-equality, rigidity, crystallization). For each ofthe eight regimes listed in Table 2, a particu-lar class of assets (e.g., economic) has beenassumed to be especially crucial in definingthe structure of inequality, thus making itpossible to specify the dominant socialclasses in terms of those assets. The rigidityof stratification systems can then be indexedby the amount of class persistence (see col-umn 5), and crystallization can be indexedby the presumed correlation between classmembership and each of the assets listed inTable 1 (see column 6).8 The final column inTable 2 rests on the further assumption thatinequality regimes have (reasonably) coher-ent ideologies that legitimate the rules andcriteria by which individuals are allocated topositions in the class structure (see column7). In most cases, ideologies of this kind are

Gloom, Doom, and Inequality – 13

Table 2. Parameters of Stratification for Eight Ideal-Typical Systems

0813343739-01.qxd 10/1/07 11:32 AM Page 13

largely conservative in their effects, but theycan sometimes serve as forces for change aswell as stability. If, for example, the facts oflabor market processes are inconsistent withthe prevailing ideology (e.g., racial discrimi-nation in advanced industrialism), then vari-ous sorts of ameliorative action might be an-ticipated (e.g., affirmative action programs).

The staple of modern classification effortshas been the tripartite distinction among class,caste, and estate, but there is also a long and il-lustrious tradition of Marxian typologicalwork that introduces the additional categoriesof primitive communism, slave society, andsocialism (e.g., Wright, Ch. 11). As shown inTable 2, these conventional approaches arelargely complementary, and it is therefore pos-sible to fashion a hybrid classification that in-corporates most of the standard distinctions.This hybrid typology should be understood asa highly stylized and compressed history of in-equality forms over the premodern, modern,and late modern periods.

The stratification forms represented inTable 2 should thus be taken as ideal types.In constructing these categories, our inten-tion is not to make empirical claims abouthow existing systems operate in practice, butrather to capture and distill the accumulatedwisdom about how these systems might op-erate in their purest form. These ideal-typicalmodels can nonetheless assist us in under-standing empirical systems. Indeed, insofaras societies evolve through the gradual over-laying of new stratification forms on older(and partly superseded) ones, it becomespossible to interpret contemporary systemsas a complex mixture of several of the idealtypes presented in Table 2.

The first panel in this table pertains to thetribal systems that dominated human societyfrom the very beginning of human evolutionuntil the Neolithic revolution of some10,000 years ago. The characterizations ofcolumns 2–7 necessarily conceal much vari-ability in the structure of these systems: It is“merely in the night of our ignorance [that]

all alien shapes take on the same hue” (An-derson, 1974, p. 549). While such variablefeatures of tribal societies are clearly of inter-est, for our purposes the important similari-ties are that (1) the total size of the distrib-utable surplus was in all cases quite limited,and (2) this cap on the surplus placed corre-sponding limits on the overall level of eco-nomic inequality (but not necessarily onother forms of inequality). It is also notewor-thy that customs such as gift exchange, foodsharing, and the like were commonly prac-ticed in tribal societies and had obvious re-distributive effects. In fact, some observers(e.g., Marx [1939] 1971) treated these soci-eties as examples of “primitive communism,”as the means of production (e.g., tools, land)were owned collectively and other types ofproperty typically were distributed evenlyamong tribal members.

This is obviously not to suggest that a per-fect equality prevailed. The more powerfulmedicine men (i.e., shamans) within tribalsocieties often secured a disproportionateshare of resources, and the tribal chief couldexert considerable influence on the politicaldecisions of the day. In most cases, theseresidual forms of power and privilege werenot directly inherited, nor were they typicallyallocated in accord with well-defined ascrip-tive traits (e.g., racial traits).9 It was only bydemonstrating superior skills in hunting,magic, or leadership that tribal memberscould secure political office or acquire statusand prestige. While meritocratic forms of al-location are often seen as prototypicallymodern, in fact they were present in incipi-ent form at the very earliest stages of societaldevelopment.

With the emergence of agrarian forms ofproduction, the economic surplus becamelarge enough to support more complex sys-tems of stratification. Among Marxist theo-rists, the “Asiatic mode” was often treated asan intermediate formation in the transitionto advanced agrarian society (e.g., feudal-ism), and we have therefore led off our typol-

14 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 14

ogy with the Asiatic case (see line B2).10 Indoing so, we should emphasize that the ex-plicit evolutionary theories of Godelier(1978) and others have not been well re-ceived, yet many scholars still take the fall-back position that Asiaticism is an importantanalytical category in the development ofclass society (e.g., Hobsbawm, 1965). Themain features of this formation are (1) a largepeasant class residing in agricultural villagesthat are “almost autarkic” (O’Leary 1989, p.17), (2) the absence of strong legal institu-tions recognizing private property rights, (3)a state elite that extracts surplus agriculturalproduction through rents or taxes and ex-pends it on “defense, opulent living, and theconstruction of public works” (Shaw 1978,p. 127),11 and (4) a constant flux in elite per-sonnel due to “wars of dynastic successionand wars of conquest by nomadic warriortribes” (O’Leary 1989, p. 18).

Beyond this skeletal outline, all else is opento dispute. There are long-standing debates,for example, about how widespread the Asi-atic mode was and about the appropriatenessof reducing all forms of Asian developmentto a “uniform residual category” (Anderson1974, pp. 548–49). These issues are clearlyworth pursuing, but for our purposes it suf-fices that the Asiatic mode provides a con-ventional example of how a “dictatorship ofofficialdom” can flourish in the absence ofprivate property and a well-developed pro-prietary class (Gouldner 1980, pp. 327–28).Under this reading of Asiaticism, the parallelwith modern socialism looms large (at leastin some quarters), so much so that variousscholars have suggested that Marx down-played the Asian case for fear of exposing itas a “parable for socialism” (see Gouldner1980, pp. 324–52).

Whereas the institution of private propertywas underdeveloped in the East, the rulingclass under Western feudalism was, by con-trast, very much a propertied one.12 The dis-tinctive feature of feudalism was that the no-bility not only owned large estates or manors

but also held legal title to the labor power ofits serfs (see line B3).13 If a serf fled to thecity, this was considered a form of theft: Theserf was stealing that portion of his or herlabor power owned by the lord (see Wright,Ch. 11). With this interpretation, the sta-tuses of serf and slave differ only in degree,and slavery thereby constitutes a limitingcase in which workers lose all control overtheir own labor power (see line B4). At thesame time, it would obviously be a mistaketo reify this distinction, given that the his-tory of agrarian Europe reveals “almost infi-nite gradations of subordination” (Bloch1961, p. 256) that confuse and blur the con-ventional dividing lines between slavery, serf-dom, and freedom. The slavery of Roman so-ciety provides the best example of completesubordination, whereas some of the slaves ofthe early feudal period were bestowed withrights of real consequence (e.g., the right tosell surplus product), and some of the (nom-inally) free men were in fact obliged to pro-vide rents or services to the manorial lord(Bloch 1961, pp. 255–74).14 The socialclasses that emerged under European agrari-anism were thus structured in quite compli-cated ways. In all cases, we nonetheless findthat property ownership was firmly estab-lished and that the life chances of individualswere defined, in large part, by their controlover property in its differing forms. Unlikethe ideal-typical Asiatic case, the nation-statewas largely peripheral to the feudal stratifica-tion system, since the means of production(i.e., land, labor) were controlled by a propri-etary class that emerged quite independentlyof the state.15

The historical record makes it clear thatagrarian stratification systems were not al-ways based on strictly hereditary forms of so-cial closure (see panel B, column 5). The caseof European feudalism is especially instruc-tive in this regard, since it suggests that strat-ification systems often become more rigid asthe underlying institutional forms matureand take shape. Although it is well known

Gloom, Doom, and Inequality – 15

0813343739-01.qxd 10/1/07 11:32 AM Page 15

that the era of classical feudalism (i.e., post-twelfth century) was characterized by a“rigid stratification of social classes” (Bloch1961, p. 325),16 there was greater perme-ability during the period prior to the institu-tionalization of the manorial system and theassociated transformation of the nobilityinto a legal class. In this transitional period,access to the nobility was not yet legally re-stricted to the offspring of nobility, nor wasmarriage across classes or estates formallyprohibited (see Bloch 1961, pp. 320–31).The case of ancient Greece provides a com-plementary example of a (relatively) openagrarian society. As Finley (1960) and othershave noted, the condition of slavery was in-deed heritable under Greek law, yet manu-mission was so common that the slave classhad to be constantly replenished with newcaptives secured through war or piracy. Thepossibility of servitude was thus somethingthat “no man, woman, or child, regardless ofstatus or wealth, could be sure to escape”(Finley 1960, p. 161). At the same time,hereditary forms of closure were more fullydeveloped in some slave systems, most no-tably the American one. As Sio (1965)notes, slavery in the antebellum South was“hereditary, endogamous, and permanent”(p. 303), with the annual manumission rateapparently as low as 0.04 percent by 1850(see Patterson 1982, p. 273). The slave soci-eties of Jamaica, South Africa, and rural Iraqwere likewise based on largely permanentslave populations.

The most extreme examples of hereditaryclosure are of course found in caste societies(see line B5). In some respects, Americanslavery might be seen as having “caste-likefeatures” (see Berreman 1981), but HinduIndia clearly provides the defining case ofcaste organization.17 The Indian caste systemis based on (1) a hierarchy of status group-ings (i.e., castes) that are ranked by ethnicpurity, wealth, and access to goods or ser-vices, (2) a corresponding set of “closure

rules” that restrict all forms of intercaste mar-riage or mobility and thereby make castemembership both hereditary and permanent,(3) a high degree of physical and occupa-tional segregation enforced by elaborate rulesand rituals governing intercaste contact, and(4) a justifying ideology that induces thepopulation to regard such extreme forms ofinequality as legitimate and appropriate(Smaje 2000; Dumont 1970; Srinivas 1962).What makes this system so distinctive, then,is not merely its well-developed closure rulesbut also the fundamentally honorific (andnoneconomic) character of the underlyingsocial hierarchy. As indicated in Table 2, thecastes of India are ranked on a continuum ofethnic and ritual purity, with the highest po-sitions in the system reserved for castes thatprohibit behaviors that are seen as dishonor-able or polluting. Under some circum-stances, castes that acquired political andeconomic power eventually advanced in thestatus hierarchy, yet they typically did so onlyafter mimicking the behaviors and lifestylesof higher castes.

The defining feature of the industrial era(see panel C) has been the emergence ofegalitarian ideologies and the consequentdelegitimation of the extreme forms of strat-ification found in caste, feudal, and slavesystems. This can be seen, for example, inthe European revolutions of the eighteenthand nineteenth centuries that pitted theegalitarian ideals of the Enlightenmentagainst the privileges of rank and the politi-cal power of the nobility. In the end, thesestruggles eliminated the last residue of feu-dal privilege, but they also made new typesof inequality and stratification possible.Under the class system that ultimatelyemerged (see line C6), the estates of the feu-dal era were replaced by purely economicgroups (i.e., “classes”), and closure rulesbased on heredity were likewise supplantedby (formally) meritocratic processes. The re-sulting classes were neither legal entities nor

16 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 16

closed status groupings, and the associatedclass-based inequalities could therefore berepresented and justified as the natural out-come of competition among individualswith differing abilities, motivation, or moralcharacter (i.e., “classical liberalism”). As in-dicated in line C6 of Table 2, the class struc-ture of early industrialism had a clear “eco-nomic base” (Kerbo 1991, p. 23), so muchso that Marx (Ch. 9) defined classes in termsof their relationship to the means of eco-nomic production. The precise contours ofthe industrial class structure are nonethelessa matter of continuing debate (as discussedabove). For example, a simple Marxianmodel focuses on the cleavage between capi-talists and workers, while more elaborateMarxian and neo-Marxian models identifyadditional intervening or “contradictory”classes (e.g., Wright, Ch. 11), and yet other(non-Marxian) approaches represent theclass structure as a continuous gradation ofwealth or income.

Whatever the relative merits of these mod-els might be, the ideology underlying the so-cialist revolutions of the nineteenth andtwentieth centuries was of course explicitlyMarxist. The intellectual heritage of theserevolutions and their legitimating ideologiescan again be traced to the Enlightenment,but the rhetoric of equality that emerged inthis period was now directed against the eco-nomic power of the capitalist class ratherthan the status and honorific privileges of thenobility. The evidence from Eastern Europeand elsewhere suggests that these egalitarianideals were only partially realized (e.g.,Lenski, Ch. 3). In the immediate postrevolu-tionary period, factories and farms were in-deed collectivized or socialized, and variousfiscal and economic reforms were institutedfor the express purpose of reducing incomeinequality and wage differentials amongmanual and nonmanual workers. Althoughthese egalitarian policies were subsequentlyweakened through the reform efforts of

Stalin and others, inequality on the scale ofprerevolutionary society was never reestab-lished among rank-and-file workers (cf.Lenski, Ch. 3). There nonetheless remainedsubstantial inequalities in power and author-ity. Most notably, the socialization of pro-ductive forces did not have the intended ef-fect of empowering workers, as the capitalistclass was replaced by a “new class” of partyofficials and managers who continued tocontrol the means of production and to allo-cate the resulting social surplus (see Eyal,Szelényi, and Townsley, Chs. 34, 104). Thisclass has been variously identified with intel-lectuals or intelligentsia, bureaucrats or man-agers, and party officials or appointees(Gouldner, Ch. 32). Regardless of the for-mulation adopted, the presumption is thatthe working class ultimately lost out in con-temporary socialist revolutions, just as it didin the so-called bourgeois revolutions of theeighteenth and nineteenth centuries.

Whereas the means of production were so-cialized in the revolutions of Eastern Europeand the former Soviet Union, the capitalistclass remained largely intact throughout theprocess of industrialization in the West. In theend, the propertied class may nonetheless beweakened by ongoing structural changes, withthe most important of these being (1) the riseof a service economy and the growing powerof the “service class” (Esping-Anderson, Ch.102; Giddens, Ch. 106), (2) the increasingcentrality of theoretical knowledge in the tran-sition to a new “information age” (Bell, Ch.101), and (3) the consequent emergence oftechnical expertise, educational degrees, andtraining certificates as “new forms of property”(Berg 1973, p. 183; Gouldner, Ch. 32;Brooks, Ch. 33). The foregoing developmentsall suggest that human and cultural capital arereplacing economic capital as the principalstratifying forces in advanced industrial soci-ety (see line C8). By this formulation, a dom-inant class of cultural elites may be emergingin the West, much as the transition to state

Gloom, Doom, and Inequality – 17

0813343739-01.qxd 10/1/07 11:32 AM Page 17

socialism (allegedly) generated a new class ofintellectuals in the East.

The Role of Benign Narratives in Past Scholarship

The foregoing classification embodies a com-plicated amalgam of “benign narratives”about how inequality has developed withinthe industrial period, narratives that are be-nign in the sense that history is understoodas operating in the main to reduce inequality,if only gradually and fitfully. Under such nar-ratives, the problem of inequality becomes atractable moral problem, an unfortunateside-circumstance of capitalism (and even so-cialism) that becomes yet more manageablewith the transition into the increasingly af-fluent forms of advanced industrialism. Thisorientation to inequality is expressed in stan-dard postwar narratives about three types ofoutcomes: (1) the distribution of income,power, and other valued resources; (2) thedistribution of opportunities for securing in-come, power, and other valued resources;and (3) the formation of social classes andother institutionalized groups (e.g., racialgroups, gender groups). We have hinted atthese narratives in the preceding section andthroughout this essay but turn in this sectionto laying them out more formally.

Trend in InequalityThe dominant inequality narrative of thepostwar period featured the emergence ofegalitarian ideologies and the consequentdelegitimation of extreme forms of inequal-ity (e.g., Kerr et al., Ch. 100; Bell, Ch. 101).The Enlightenment is understood in thiscontext as unleashing one of the most pro-found revolutions in human history. Theresulting decline in inequality can be seen,for example, in (1) the European revolu-tions of the eighteenth and nineteenth cen-turies against the privileges of rank (hon-orific equality); (2) the gradual eliminationof inequalities in the right to vote, own

property, and speak and assemble (civilequality); (3) the abolition of slavery andthe establishment of the radically egalitar-ian principle of self-ownership (equality ofhuman assets); and (4) the equalization ofeconomic assets via the rise of socialism,welfare capitalism, and their many institu-tions (economic equality).

As is well known, the latter commitmentto equalizing economic assets was ratherweaker than the commitment to other formsof equalization (at least in the capitalist con-text), with the result that economic inequali-ties remained extreme in all marketeconomies. There was nonetheless a gradualdecline in economic inequality throughoutthe postwar period in the United States andother industrial countries. According to theclassic Kuznets curve (Kuznets 1955), theinitial stages of capitalist development bringabout an increase in income inequality ascapital is increasingly concentrated among asmall number of investors, whereas more ad-vanced forms of capitalism entail a growth inthe size of the middle class and a consequentreversal of the upward trend. The causal dy-namics behind the resulting inverted-U pat-tern remain unclear (see Piketty and Saez,Ch. 8), but most sociologists attribute thelate-industrial decline in inequality to the in-creasingly crucial role that the skilled work-ing class played in production, the associatedgrowth in working-class productivity, andthe leverage that this growth in skills andproductivity conferred on skilled workers.

It is worth inquiring as to the less proxi-mate mechanisms by which egalitarianismof this sort may diffuse and take hold. Theconventional view in this regard is that a se-ries of crucial historical events after the En-lightenment (e.g., the defeat of Nazism, thecivil rights movement) served to defineequality as one of our core cultural commit-ments. Absent some revolutionary event thatchanges this cultural trajectory, the course ofhuman history then becomes the “workingout” of this commitment, a task that in-

18 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 18

volves shedding subsidiary values that some-times come into conflict with our deepercommitment to egalitarianism. The coremechanism that drives this cultural diffusionmay therefore be understood as the gradualreconciling of competing values to a newvalue, that of equality, that has been elevatedby one or more historical events to a posi-tion of prominence.

This is obviously not to suggest that allpostwar sociologists and social scientists em-phasized forces making for a decline in in-equality. However, even when a benign nar-rative was not explicitly adopted, there wasusually some effort to engage with it and toexplain or defend the decision not to take itup. This defensiveness was especially appar-ent in neo-Marxian analyses of the postwarera. Although such analyses were based ondeeply pessimistic subnarratives about thetrajectory of capitalism, these subnarrativeswere typically attached to larger and morebenign narratives about the postcapitalist tra-jectory (e.g., Marx, Ch. 9; also see Wright,Ch. 11; Wallerstein, Ch. 12).

Trend in Inequality of OpportunityThe second benign narrative of interest restson the conventional distinction between thedistribution of social rewards (e.g., income)and the distribution of opportunities for se-curing these rewards. In liberal welfareregimes, extreme inequalities in rewards maybe tolerated, but only insofar as opportuni-ties for attaining these rewards are under-stood to be equally distributed. It is inequal-ities of opportunity that are regarded, then,as especially illegitimate in the context of lib-eral welfare regimes.

The dominant narratives of the postwarperiod have these inequalities of opportunitygradually weakening. The narratives of thisperiod may be understood as benign becausethey describe the withering away of preciselythose types of inequalities (i.e., inequalitiesof opportunity) that are regarded as prob-lematic or illegitimate. The trademark of the

benign narrative is this simple correspon-dence between what we want and what wethink will likely happen. We describe belowfour benign subnarratives that characterizesome of the processes by which inequalitiesof opportunity may come to be weakened.

The most famous such subnarrative per-tains to the discrimination-reducing effectsof competitive market economies. In hisoriginal formulation of the “taste for discrim-ination” model, Becker (1957) argued thatdiscrimination would be eroded by competi-tive market forces because it requires em-ployers to pay a premium to hire members ofthe preferred class of labor, whether these bemales, whites, or any other ascriptively de-fined classes. This taste is “discriminatory”because it rests on exogenous preferences fora certain category of labor that cannot be un-derstood as arising from some larger concernfor maximizing profitability or market share.When managers make hiring decisions in ac-cord with such tastes, their firms will not becompetitive with nondiscriminating firmsbecause they must pay extra to secure laborfrom the preferred class (without any com-pensating increase in productivity). In stan-dard renditions of this account, it is pre-sumed that discriminating firms willgradually be selected out by the market, al-though it is also possible that some discrimi-nating firms will change their hiring prac-tices to remain competitive.

This economic subnarrative works in tan-dem with a second, “organizational” one thatemphasizes the diffusion of modern person-nel practices in the form of universalistic hir-ing practices (e.g., open hiring, credential-ism) and bureaucratized pay scales andpromotion procedures (Weber, Ch. 13;Grusky and Levanon, Ch. 87). The essenceof such bureaucratic personnel practices is aformal commitment to universalism (i.e.,treating all workers equally) and to merito-cratic hiring and promotion (i.e., hiring andpromoting on the basis of credentials). In itsideal-typical form, the spread of bureaucracy

Gloom, Doom, and Inequality – 19

0813343739-01.qxd 10/1/07 11:32 AM Page 19

becomes an organizational process with itsown dynamic, a process of diffusion thatrests not on actual efficiencies, as with theeconomic subnarrative, but simply on thepresumption that bureaucratic practices areefficient and that “modern firms” must there-fore adopt them. This subnarrative, like theeconomic one, implies that firms will gradu-ally come to embrace organizational proce-dures that reduce inequalities of opportunity.

The third subnarrative of interest is thepolitical one. Whereas the economic andorganizational subnarratives treat change ininequality as an unintended by-product ofmacro-level forces (i.e., competition andbureaucratization), the political subnarra-tive is about instrumental action explicitlyoriented towards effecting a decline in in-equality. In theory, such political actioncould be oriented toward reducing eitherinequalities of opportunity or outcome,but historically a main emphasis within lib-eral welfare regimes has been legislationaimed at reducing inequality of opportu-nity (e.g., antidiscrimination legislation,early education programs, educationalloans). The distinctive assumption of thepolitical subnarrative is that straightfor-ward “social engineering” is an importantsource of change and that the unintendedor unanticipated consequences of such en-gineering are too often overemphasized.

The final subnarrative, a simple culturalone, rests on the argument that Westernideals of justice and equality continue to beendogenously worked out through a logicthat diffuses independently of the economicefficiency of such ideals. The cultural subnar-rative can be straightforwardly distinguishedfrom the economic one because the growing“taste” for equality is presumed to be an ex-ogenous shift rather than some accommoda-tion to the economic cost of exercising dis-criminatory tastes. Likewise, the culturalsubnarrative is distinct from the organiza-tional subnarrative by virtue of focusing onthe spread of tastes for equality and equality-

enhancing practices, not the spread of orga-nizational forms (e.g., bureaucratization)that are deemed efficient, normatively desir-able, or both. Similarly, the cultural and po-litical subnarratives are closely related be-cause political commitments to equalopportunity, antidiscrimination legislation,and school reform may be partly or evenlargely motivated by these newfound tastesfor equal opportunity. At the same time, thecultural commitment to equal opportunity isnot expressed exclusively in such politicalterms but is additionally expressed in the at-titudes, behaviors, and personnel practices ofemployers. Most obviously, employers maygradually shed their preferences for certaincategories of labor and instead develop posi-tive tastes for equality in hiring, firing, andpromotion, tastes that might at the limit beexercised in the labor market even with someloss in profits or efficiency.

The spread of such tastes for equal oppor-tunity may again be viewed as part of ourEnlightenment legacy, albeit a particular“liberal” variant of that legacy that empha-sizes equalizing opportunities, not outcomes.This commitment is expressed not only atthe individual level (e.g., changes in atti-tudes) but also at the collective level throughvarious types of political reform (e.g., an-tidiscrimination legislation) as well as thediffusion of bureaucratic personnel policies(e.g., open hiring).

Trend in Class FormationThe final benign narrative of interest de-scribes the gradual transition from “lumpy”class-based labor markets to more purelygradational ones (see Weeden et al., Ch. 27).Within this narrative, the early-industrialeconomy is represented as deeply balkanizedinto partly independent labor markets de-fined by detailed occupations (e.g., econo-mist, carpenter), big social classes (e.g.,manager, farmer), or yet more aggregatedfactors of production (e.g., worker, capital-ist). For our purposes, what is principally of

20 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 20

interest is our collective fascination with ar-guments describing how these classes, how-ever they may be defined, tend to graduallydissipate and leave us with gradational labormarkets that increasingly approximate theseamless neoclassical ideal. The first step inthis transition, as described most famouslyby Dahrendorf (Ch. 10), is the gradual “in-stitutionalization” of class conflict, a regular-ization of labor-capital relations achievedthrough the establishment of unions, collec-tive bargaining agreements, and other lawsdefining how labor and capital should nego-tiate. The second step in this transition in-volves the dismantling of unions and otherinstitutionalized residues of classes as the lib-eral welfare ideals of “deregulation” and flex-ibility are increasingly pursued (Sørensen,Ch. 25). Although the mechanisms differ,there are of course analogous narratives onoffer pertaining to the gradual decline of sta-tus groupings based on race, ethnicity, orgender (e.g., Wilson, Ch. 73; also see Pe-tersen, Ch. 83).

This line of argumentation is additionallyexpressed in postmodernist narratives thatassume that class identities, ideologies, andorganization are attenuating and that “newtheories, perhaps more cultural than struc-tural, [are] in order” (Davis 1982, p. 585).The core claim here is that politics,lifestyles, and consumption practices are nolonger class-determined and increasinglybecome a “function of individual taste,choice, and commitment” (Crook, Pakul-ski, and Waters 1992, p. 222; Pakulski andWaters, Ch. 107). In more ambitious vari-ants of postmodernism, the focus shiftsaway from simply claiming that attitudesand practices are less class-determined, andthe older class-analytic objective of under-standing macro-level stratificational changeis resuscitated. This ambition underlies, forexample, all forms of postmodernism thatseek to represent “new social movements”(e.g., environmentalism) as the vanguardforce behind future stratificatory change. As

argued by Eyerman (1992) and others (e.g.,Pakulski and Waters, Ch. 107), the labormovement can be seen as a fading enterpriserooted in the old conflicts of the workplaceand industrial capitalism, whereas new so-cial movements provide a more appealingcall for collective action by virtue of theiremphasis on issues of lifestyle, personalidentity, and normative change. With thisformulation, the proletariat is stripped of itsprivileged status as a universal class, andnew social movements emerge as an alterna-tive and far more benign force “shaping thefuture of modern societies” (Haferkampand Smelser 1992, p. 17).

New Approaches to Studying Inequality

The foregoing narratives, all of which werefixtures of the postwar intellectual landscape,describe the emergence of a world in whichinequalities are less profound, opportunitiesare more equally distributed, and class con-flicts and interclass differences become atten-uated. These narratives are benign in thesense that they push us toward equilibria thatmost commentators, even neo-Marxian ones,might well regard as appealing. The benignnarrative is accordingly built on the happycorrespondence between what should be andwhat will be.

If there is any theme to contemporaryanalyses of inequality, it is that the benignnarrative has fallen largely out of fashion. Wehave nonetheless laid out these standard nar-ratives in some detail because they provide animportant backdrop to current theorizingand are often used as foils by contemporaryscholars seeking to motivate their own analy-ses. The benign narrative is in this sense lurk-ing in the background of contemporary dis-cussions of inequality. We turn now to acloser discussion of how contemporaryanalyses of inequality have developed partlyin reaction to the benign narratives of thepostwar period.

Gloom, Doom, and Inequality – 21

0813343739-01.qxd 10/1/07 11:32 AM Page 21

Multidimensionalism and New InequalitiesAs a natural starting point for this discus-sion, we note that contemporary inequalityscholarship is increasingly concerned withnew forms of inequality, forms that were ei-ther ignored in the past or have been spawnedby new technologies or institutions. Thisgrowing emphasis on new inequalities isconsistent with the now fashionable viewthat inequality is multidimensional andthat conventional studies of economic, so-cioeconomic, or cultural inequality hardlyexhaust its many forms.

This approach may be understood as sim-ply reemphasizing the importance of study-ing all eight forms of inequality listed inTable 1. If the resulting multidimensionalspace is taken seriously, we can then askhow individuals are distributed among theless conventionally studied dimensions ofthe inequality space. Are “new” assets as un-equally distributed as old ones? Is inequalitybecoming more of an “all or nothing” affairin which upper-class workers are advan-taged on all dimensions of interest andlower-class workers are disadvantaged on alldimensions of interest? Are new assetssometimes distributed in ways that com-pensate for shortfalls in older ones? In thepresent volume, multidimensionalist ques-tions of this kind are posed for such “new”outcomes as health (Mullahy, Robert, andWolfe, Ch. 95), computer literacy (Hargit-tai, Ch. 98), imprisonment or capital pun-ishment (Western, Ch. 41), and networksand social capital (Granovetter, Ch. 59; Lin,Ch. 60; Burt, Ch. 61). These new types ofinequality may be understood in some casesas truly new divides generated by new tech-nologies (e.g., the digital divide) or new so-cial institutions (e.g., modern mass pris-ons). More typically, the “new” outcomesare just increasingly popular topics of studyamong academics, not truly new forms(e.g., health inequalities).

The Intransigence of Poverty andInequalityThe foregoing line of research typically takesthe form of an exposé of the extent to whichseemingly basic human entitlements, such asliving outside of prison, freely participatingin digital culture, or living a long and healthylife, are unequally distributed in ways thatsometimes amplify well-known differentialsof income or education. The continuing at-traction of such exposés (at least among aca-demics) may be attributed to our collectivediscomfort with an economic system thatgenerates rather more inequality than ispalatable under contemporary cultural stan-dards. Although the equalizing reforms of so-cial democracy have historically been a mainsolution to this tension, the declining legiti-macy of such reform (especially in Europeand the United States) leaves the tension anincreasingly unresolved one.

Whereas the old narratives focused, then,on the forces making for decline in inequal-ity, a more pessimistic assessment of the tra-jectory of late industrialism has now takenhold, and much scholarship accordingly fo-cuses on documenting that inequality haspersisted at higher levels than had been an-ticipated. This sensibility underlies, for ex-ample, contemporary research showing thatresidential segregation in the United States isso extreme as to constitute a modern form of“apartheid” (Massey and Denton, Ch. 38),that racial discrimination in labor marketslikewise remains extreme (Bertrand and Mul-lainathan, Ch. 70; Pager, Ch. 72), that theoccupational structure is “hypersegregated”by gender (Reskin, Ch. 86), that income in-equality has increased markedly in manycountries over the last thirty-five years(Piketty and Saez, Ch. 8), that poverty ratesin the United States remain strikingly high(Smeeding, Ch. 36), that African Americansare routinely harassed, slighted, and insultedin public places (Feagin, Ch. 74), that work-ing-class and middle-class children tend to

22 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 22

be raised in profoundly different ways(Lareau, Ch. 97), that political behavior con-tinues to be strongly shaped by class-basedpolitics (Hout and Moodie, Ch. 99), andthat massive class disparities in access tohealth services persist (Mullahy, Robert, andWolfe, Ch. 95).

The cynic might ask whether this newmuckraking tradition is really all that neces-sary. Is there truly a large public that doesn’talready appreciate the persistence of many ofthese inequalities? We live, after all, in a mar-ket society in which virtually everything iscommodified, meaning that almost all goodsand services (e.g., health care, housing) areallocated on the basis of our ability to pay forthem. Because we are so deeply and (seem-ingly) irrevocably marketized, the real intel-lectual challenge would be to find a good,service, or outcome that is somehow un-touched by class, one that is perfectly andequally distributed to all. The role of class isin this sense so obvious as to preclude anyneed for academic muckraking.

We suspect that our contributors wouldreact to such (hypothetical) criticism by em-phasizing that only some inequalities maybe understood as the inevitable outcome ofour collective decision to allocate resourceson the basis of a market. It is at least possi-ble to imagine markets that do not entailracial discrimination, that do not entailgender discrimination, and that encompassinstitutions that have a substantial inequal-ity-moderating effect. It is likewise possibleto imagine that class differences in politics,culture, and child rearing practices wouldhave by now abated. The benign narrativesof the postwar period in fact laid out pre-cisely such imaginings. If it is now clear thatthese imagined futures have not been real-ized, surely we need to document that con-clusion with all the rigor that can possiblybe mustered. To be sure, many of us wellknow that the world is a massively unequalone, but even so the force of the known can

be readily lost when we live with profoundinequality on a day-to-day basis. This com-mitment to remind us of what comes to betaken for granted is the cornerstone, we sus-pect, of the renewed interest in the inequality-documenting function.

The Rise of Less Benign NarrativesThe rise of this muckraking exposé of in-equality has been coupled, moreover, withincreasing interest in developing narrativesthat explain why inequality has persisted orgrown more extreme. These narratives aretypically less grand than the quite encom-passing narratives of the postwar period; thatis, rather specialized narratives have recentlydeveloped around many of the various unittrends of interest (e.g., the expansion of in-come inequality, the stalling decline in thegender wage gap), and rather little attentionhas been paid to developing some grandmeta-narrative that links these specializednarratives together. The signature, then, ofthe contemporary narrative is this highly de-limited focus, a commitment to developing arigorously empirical foundation, and a spe-cial interest in identifying those more insidi-ous social forces that undermine the benignnarratives of the past.

By way of example, consider the historicrise in income inequality, a development thathas spawned one of the most sustained ef-forts at narrative-building of our time (seeMorris and Western, Ch. 7). As noted above,the classic Kuznets curve aligns nicely withthe facts of inequality up to the early 1970s,but then a dramatic, unprecedented upswingin inequality in the post-1970 period made itclear that inequality history had not ended.We have since witnessed one of the mostmassive research efforts in the history of so-cial science as scholars sought to identify the“smoking gun” that accounted for this dra-matic increase in inequality.

Initially, the dominant hypothesis was thatdeindustrialization (i.e., the relocation of

Gloom, Doom, and Inequality – 23

0813343739-01.qxd 10/1/07 11:32 AM Page 23

manufacturing jobs to offshore labor mar-kets) brought about a decline in demand forless-educated manufacturing workers, a de-cline that generated increases in inequality byhollowing out the middle class and sendingmanufacturing workers into unemploymentor into the ranks of poorly paid service work.Although this line of argumentation still hasits advocates, it cannot easily be reconciledwith evidence suggesting that the computeri-zation of the workplace and related techno-logical change has been at least one forcebehind a heightened demand for highly edu-cated workers. Because of this result (andother supporting evidence), the deindustrial-ization story has now been largely supplantedby the converse hypothesis that “skill-biasedtechnological change” has increased the de-mand for high-skill workers beyond the in-crease in supply, thus inducing a short-termdisequilibrium and a correspondingly in-creased payoff for high-skill labor. At thesame time, most scholars acknowledge thatthis story is at best an incomplete one andthat other accounts, especially more nar-rowly political ones, must additionally beentertained (e.g., Card and DiNardo 2002).Most notably, some of the rise in income in-equality in the United States was clearly at-tributable to the declining minimum wage(in real dollars), a decline that in turn has tobe understood as the outcome of politicalprocesses that increasingly favor pro-in-equality forces (Morris and Western, Ch. 7).The same conclusion applies yet more obvi-ously to the recent round of tax cuts in theUnited States.

The future of income inequality dependson which of these underlying mechanisms isprincipally at work. The silver lining of thedeindustrialization story is that within-coun-try increases in inequality should be offset bybetween-country declines (as poor countriesprofit from new manufacturing jobs),whereas the silver lining of skill-biased tech-nological change is that the heightened de-mand for high-skill workers is presumably a

one-time, short-term disequilibrium thatwill, by virtue of the higher payoff to high-skill jobs, trigger a compensating growth inthe supply of high-skill workers. There is,unfortunately, no shortage of competing sto-ries that imply more disturbing futures, evenfutures consistent with a classical Marxianaccount in which low-skill workers are emis-erated within some countries by virtue of aglobalization-induced “race to the bottom.”Indeed, accounts that focus on the politicalsources of rising inequality often take on thismore disturbing character, given that socialdemocratic ideologies have fallen largely outof fashion and no longer provide capitalistswith a viable high road of “enlightened self-interest” (e.g., support for labor unions, re-distribution). As social democratic agendascome to be viewed with suspicion, politicalsupport for the minimum wage and other in-equality-reducing institutions may increas-ingly falter, and market-generated inequalitymay no longer be much restrained by pre-market or after-market interventions.

We have focused on the rise of income in-equality and the various narratives it has gen-erated only because this literature is espe-cially well known and central to the field.The larger point that we seek to make is that,no matter the subfield, there appears to bemuch interest in developing narratives thatexplain why long-standing declines in in-equality have slowed down, stalled alto-gether, or even reversed themselves. We arereferring, for example, to (1) narratives of“globalization” that describe how the liberal-ization of financial and capital markets hasharmed poor countries (Stiglitz, Ch. 109);(2) narratives of “deindustrialization” thatdescribe the loss of inner-city jobs and the as-sociated rise of an urban underclass (Ehren-reich, Ch. 35; Wilson, Ch. 37); (3) narrativesof “deunionization” that describe the loss ofmiddle-class unionized jobs and the emer-gence of nonstandard forms of employment(e.g., Piore, Ch. 56; Kalleberg, Ch. 58); (4)narratives of “segmented assimilation” that

24 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 24

describe the relatively bleak prospects for atleast some new immigrant groups (Portesand Zhou, Ch. 68; Waters, Ch. 69); (5) nar-ratives of “opting out” that have highlytrained women eschewing stressful careersin favor of recommitting to their children,spouses, and domestic responsibilities (Belkin,Ch. 78; cf. Boushey, Ch. 79); (6) narrativesof “essentialist segregation” that describe howsex-typed occupational ghettos continue tobe built around presumed differences in maleand female aptitudes (Grusky and Levanon,Ch. 87), (7) narratives of “nonconscious dis-crimination” that refer to subtle internalizedprejudices that are especially difficult to ex-tirpate (e.g., Correll, Benard, and Paik, Ch.81; Reskin, Ch. 82); and (8) narratives of“spatial segregation” that emphasize thedeeply institutionalized forces generatingracial enclaves (Massey and Denton, Ch. 38)and poverty-stricken neighborhoods (Pebleyand Sastry, Ch. 39; Sampson and Morenoff,Ch. 40). The key question of our time, andone which remains largely unanswered bythe evidence of the last half-century, iswhether the forces for equality featured inthe “liberal theory” are strong enough toovercome the above countervailing processes.Although counternarratives of the more opti-mistic sort are also being developed (e.g.,Firebaugh, Ch. 110), these seem not to be asfrequently generated or as readily embraced,and the proponents of such narratives tendnow to find themselves beleaguered, out-numbered, and on the defensive.

Conclusions

We have fashioned the latter part of our re-view around the subtle changes over the lasthalf-century in the types of narratives that so-ciologists and other social scientists have ap-plied to make sense of trends in inequality.The narratives of the postwar period took ona strikingly benign form in which the domi-nant logics of history were understood as op-erating in the main to reduce inequality. These

benign narratives, which now mainly seemnaive and quaint, have been supplanted by ahost of new narratives that give far greaterweight to the forces making for inequality.

Has the pendulum swung too far? It is ofcourse child’s play to posit any number ofnonempirical sources of our fascination,some might say obsession, with the pes-simistic narrative. The following are perhapsthe most obvious of such accounts:

The Newsworthiness AccountIf the postwar era of the benign narrativewas indeed one of across-the-board declinesin inequality, then a special incentive pre-sumably emerged to ferret out results thatwere inconsistent with the prevailing wis-dom and hence would be regarded as news-worthy. It is surely difficult to market analy-ses and forge careers on the basis ofbusiness-as-usual evidence. We might ask,for example, whether the “digital divide”emerged as a newsworthy topic because ofearly evidence of a substantial divide. If in-stead access to computing was found to beequal, would we have quickly discarded thetopic and set off to find some other moreunequal outcome? This type of selection onthe dependent variable (i.e., inequality) willcreate a research literature that exaggerateshow unequal the world is.

The “Moral Credentials” AccountIt is well to bear in mind that the contempo-rary academic, far from taking a vow ofpoverty, is now firmly ensconced in the mid-dle class, often the upper middle class. Whenrelatively privileged scholars study povertyand disadvantage, they often feel a specialobligation to demonstrate a strong commit-ment to amelioration and to display politicalsentiments that are liberal, progressive, oreven radical. In some cases, this pressure maymotivate them to downplay any evidence ofdecline in inequality or disadvantage, pre-sumably out of concern that undue emphasison the progress achieved so far will make it

Gloom, Doom, and Inequality – 25

0813343739-01.qxd 10/1/07 11:32 AM Page 25

appear that the remaining disadvantage istaken as acceptable or unproblematic.

The “Obsession with SmallDifferences” AccountThe continuing diffusion of egalitarian val-ues renders any departures from equality, nomatter how small, as problematic and news-worthy. By this logic, even increasingly smallintergroup differences will attract much at-tention, especially because ever more power-ful models and statistical methods now makeit possible to tease them out.

If these nonempirical interpretationsshouldn’t be dismissed out of hand, norshould they be taken too seriously. It is rele-vant in this regard that many of the pes-simistic narratives involve indisputably dra-matic changes in outcomes (e.g., income)that were frequently studied well before anyreversal in the trendline was detected. Al-though some of the pessimistic narrativesfeatured in this volume are still in incipientform and are not yet well researched, it willof course be difficult to continue to maintainthese narratives should strong disconfirmingevidence turn up. The implication, then, isthat our exaggerated taste for pessimismmight conceivably lead us to cycle through agreat many pessimistic stories (and fail to de-velop enough benign ones), but at least theusual rules of evidence will have us exciseegregiously flawed narratives, no matter howbenign or pessimistic they may be.

NOTES1. In fact, the term stratification has itself been

seen as anti-Marxist by some commentators (e.g.,Duncan 1968), since it places emphasis on the verti-cal ranking of classes rather than the exploitative re-lations between them. The geological metaphor im-plied by this term does indeed call attention to issuesof hierarchy. Nonetheless, whenever it is used in thepresent essay, the intention is to refer generically toinequality of all forms.

2. In some stratification systems, the distributionof rewards can be described with a single matchingalgorithm, since individuals receive rewards directly

rather than by virtue of the social positions thatthey occupy. The limiting case here would be thetribal economies of Melanesia in which “Big Men”(Oliver 1955) secured prestige and power throughpersonal influence rather than through incumbencyof any well-defined roles (also see Granovetter1981, pp. 12–14).

3. It goes without saying that the assets listed inTable 1 are institutionalized in quite diverse ways.For example, some assets are legally recognized bythe state or by professional associations (e.g., civilrights, property ownership, educational credentials),others are reserved for incumbents of specified workroles (e.g., workplace authority), and yet others haveno formal legal or institutional standing and are re-vealed probabilistically through patterns of behaviorand action (e.g., high-status consumption practices,deference, derogation).

4. It is sometimes claimed that educational cre-dentials are entirely investment goods and shouldtherefore be excluded from any listing of the primi-tive dimensions underlying stratification systems(e.g., Runciman 1968, p. 33). In evaluating thisclaim, it is worth noting that an investment rhetoricfor schooling became fashionable only quite recently(e.g., Becker 1975), whereas intellectuals and hu-manists have long viewed education as a simple con-sumption good.

5. This is not to gainsay the equally importantpoint that parents often encourage their children toacquire such goods because of their putative benefits.

6. Although “native ability” is by definition estab-lished at birth, it is often seen as a legitimate basisfor allocating rewards (because it is presumed to berelevant to judgments of merit).

7. Although Pakulski and Waters (Ch. 107) usethe label postmodern in their analyses, other scholarshave invented such alternative terms as late moder-nity, high modernity, or reflexive modernization (seeGiddens, Ch. 106; Beck and Lau, Ch. 108), and yetothers continue to use modernity on the argumentthat the changes at issue are mere extensions of thoselong under way. We frequently use the term post-modern in this essay without intending to disadvan-tage the analyses of those who prefer other labels.

8. The assumptions embedded in columns 4–6 ofTable 2 are clearly far-reaching. Unless a stratifica-tion system is perfectly crystallized, its parametersfor inequality and rigidity cannot be represented asscalar quantities, nor can the intercorrelations be-tween the multiple stratification dimensions be eas-ily summarized in a single parameter. Moreover,even in stratification systems that are perfectly crys-tallized, there is no reason to believe that persistence

26 – Melvin M. Tumin

0813343739-01.qxd 10/1/07 11:32 AM Page 26

over the life course (i.e., intragenerational persis-tence) will always vary in tandem with persistencebetween generations (i.e., intergenerational inheri-tance). We have nonetheless assumed that each ofour ideal-typical stratification systems can be charac-terized in terms of a single “rigidity parameter” (seecolumn 5).

9. This claim does not hold with respect to gen-der. That is, women and men were typically assignedto different roles, which led to consequent differ-ences in the distribution of rewards.

10. It should again be stressed that our typologyby no means exhausts the variability of agrarianstratification forms (see Kerbo 2000 for an extendedreview).

11. The state elite was charged with constructingand maintaining the massive irrigation systems thatmade agriculture possible in regions such as China,India, and the Middle East (cf. Anderson 1974, pp.490–92).

12. This is not to suggest that feudalism canonly be found in the West or that the so-calledAsiatic mode is limited to the East. Indeed, the so-cial structure of Japan was essentially feudalisticuntil the mid-nineteenth century (with the rise ofthe Meiji State), and the Asiatic mode has beendiscovered in areas as diverse as Africa, pre-Columbian America, and even Mediterranean Eu-rope (see Godelier 1978). The latter “discoveries”were of course predicated on a broad and ahistori-cal definition of the underlying ideal type. As al-ways, there is a tension between scholars who seekto construct ideal types that are closely tied to his-torical social systems and those who seek to con-struct ones that are broader and more encompass-ing in their coverage.

13. This economic interpretation of feudalism isclearly not favored by all scholars. For example,Bloch (1961, pp. 288–89) argues that the definingfeature of feudalism is the monopolization of author-ity by a small group of nobles, with the economicconcomitants of this authority (e.g., land ownership)thus being reduced to a position of secondary im-portance. The “authority classes” that emerge underhis specification might be seen as feudal analogues tothe social classes that Dahrendorf (Ch. 10) posits forthe capitalist case.

14. In the so-called secondary stage of feudalism(Bloch 1961), the obligations of serfs and free menbecame somewhat more formalized and standard-ized, yet regional variations of various sorts stillpersisted.

15. It was not until the early fourteenth centurythat states of the modern sort appeared in Europe.

16. In describing this period of classical feudal-ism, Bloch (1961) noted that “access to the circle ofknights . . . was not absolutely closed, [yet] the doorwas nevertheless only very slightly ajar” (p. 325).

17. The Indian caste system flourished during theagrarian period, yet it persists in attenuated formwithin modern industrialized India.

REFERENCESAnderson, Perry. 1974. Lineages of the Absolutist

State. London: Verso.Becker, Gary S. 1957. The Economics of Discrimina-

tion. Chicago: University of Chicago Press.______. 1975. Human Capital. Chicago: University

of Chicago Press.Berg, Ivar. 1973. Education and Jobs: The Great

Training Robbery. Harmondsworth, UK: PenguinBooks.

Berreman, Gerald. 1981. Caste and Other Inequities.Delhi: Manohar.

Bloch, Marc. 1961. Feudal Society. London: Rout-ledge and Kegan Paul.

Bottomore, Thomas B. 1965. Classes in Modern So-ciety. London: Allen and Unwin.

Card, D., and J. E. DiNardo. 2002. “Skill-BiasedTechnological Change and Rising Wage Inequal-ity: Some Problems and Puzzles.” Journal of LaborEconomics 20: 733–83.

Crook, Stephen, Jan Pakulski, and Malcolm Waters.1992. Postmodernization. London: Sage Publica-tions.

Davis, James. 1982. “Achievement Variables andClass Cultures: Family, Schooling, Job, andForty-Nine Dependent Variables in the Cumu-lative GSS.” American Sociological Review 47:569-86.

Dumont, Louis. 1970. Homo Hierarchicus. Chicago:University of Chicago Press.

Duncan, Otis Dudley. 1968. “Social Stratificationand Mobility: Problems in the Measurement ofTrend.” Pp. 675–719 in Indicators of SocialChange, edited by Eleanor B. Sheldon andWilbert E. Moore. New York: Russell Sage Foun-dation.

Eyerman, Ron. 1992. “Modernity and Social Move-ments.” Pp. 37-54 in Social Change and Moder-nity, edited by Hans Haferkamp and Neil J.Smelser. Berkeley: University of California Press.

Finley, Moses J. 1960. Slavery in Classical Antiquity.Cambridge, UK: W. Heffer and Sons.

Godelier, Maurice. 1978. “The Concept of the ‘Asi-atic Mode of Production’ and Marxist Models ofSocial Evolution.” Pp. 209–57 in Relations of Pro-

Gloom, Doom, and Inequality – 27

0813343739-01.qxd 10/1/07 11:32 AM Page 27

28 – Melvin M. Tumin

duction, Marxist Approaches to Economic Anthro-pology, edited by D. Seddon. London: Cass.

Gouldner, Alvin. 1980. The Two Marxisms: Contra-dictions and Anomalies in the Development of The-ory. New York: Seabury Press.

Granovetter, Mark. 1981. “Toward a SociologicalTheory of Income Differences.” Pp. 11–47 in So-ciological Perspectives on Labor Markets, edited byIvar Berg. New York: Academic Press.

Haferkamp, Hans, and Neil J. Smelser. 1992. “In-troduction.” Pp. 1-33 in Social Change andModernity, edited by Hans Haferkamp and Neil J.Smelser. Berkeley: University of California Press.

Hobsbawm, Eric J. 1965. “Introduction.” Pp. 9–65in Pre-Capitalist Economic Formations, translatedby Jack Cohen and edited by Eric J. Hobsbawm.New York: International Publishers.

Kerbo, Harold R. 1991. Social Stratification and In-equality: Class Conflict in Historical and Compara-tive Perspective. New York: McGraw-Hill.

______. 2000. Social Stratification and Inequality:Class Conflict in Historical, Comparative, andGlobal Perspective. New York: McGraw-Hill.

Kuznets, Simon. 1955. “Economic Growth and In-come Inequality (Presidential Address).” AmericanEconomic Review 45: 1–28.

Lipset, Seymour M. 1968. “Social Class.” Pp.296–316 in International Encyclopedia of the Social

Sciences, edited by D. L. Sills. New York: Macmil-lan.

Marx, Karl. [1939] 1971. The Grundrisse. New York:Harper and Row.

O’Leary, Brendan. 1989. The Asiatic Mode of Pro-duction. Oxford: Basil Blackwell.

Oliver, Douglas L. 1955. A Solomon Island Society.Cambridge: Harvard University Press.

Patterson, Orlando. 1982. Slavery and Social Death.Cambridge: Harvard University Press.

Runciman, Walter G. 1968. “Class, Status, andPower?” Pp. 25–61 in Social Stratification, editedby J. A. Jackson. New York: Cambridge Univer-sity Press.

Schumpeter, Joseph. 1953. Aufsätze zur Soziologie.Tübingen: Mohr/Siebeck.

Shaw, William H. 1978. Marx’s Theory of History.Stanford, CA: Stanford University Press.

Sio, Arnold A. 1965. “Interpretations of Slavery:The Slave Status in the Americas.” ComparativeStudies in Society and History 7: 289–308.

Smaje, Chris. 2000. Natural Hierarchies: The Histor-ical Sociology of Race and Caste. Oxford: BasilBlackwell.

Srinivas, Mysore N. 1962. Caste in Modern India.London: Asia Publishing House.

Tawney, R. H. 1931. Equality. London: GeorgeAllen and Unwin.

0813343739-01.qxd 10/1/07 11:32 AM Page 28